1. School of Computer and Information Technology, Zhejiang Changzheng Vocational and Technical College, Hangzhou 310023, China 2. College of Electrical Engineering, Zhejiang University, Hangzhou 310058, China
A novel identity recognition method based on multi-dimensional features extracted from photoplethysmography (PPG) signals was proposed, addressing the limitations of existing methods in terms of incomplete feature representation and weak robustness. The non-linear dimension of PPG signals was incorporated as a crucial feature for identity recognition. After preprocessing the PPG signals were processed, and features were extracted from three distinct dimensions, i.e., time domain, frequency domain, and non-linearity. An effective feature set was then constructed through optimization and selection. Finally, this feature set was utilized for identity recognition, and the performance of the recognition system was analyzed and evaluated. By comprehensively analyzing multiple dimensions, the proposed method achieved comprehensive feature extraction. Furthermore, the complementary information provided by time domain, frequency domain, and non-linearity analysis enhanced the robustness of the recognition system. On an identity recognition task involving 200 subjects and 1000 samples, the proposed approach achieved an accuracy of 98.4%. Comparative analysis with other state-of-the-art methods, such as KNN, demonstrated the superior accuracy of the proposed approach. Results indicated the significance of constructing multi-dimensional features for enhancing the accuracy of PPG identity recognition tasks.
Youping FU,Hang ZHANG,Menghan LI,Jun MENG. Identity recognition based on multi-dimensional features of pulse wave signals. Journal of ZheJiang University (Engineering Science), 2025, 59(3): 566-576.
Tab.1Names, definitions, and calculation formulas of time-domain characteristic parameters for PPG signals
Fig.4Time domain characteristic parameter map
参数名称
FD/Hz
参数名称
FD/Hz
power_1
0~0.02
power_5
0.13~0.16
power_2
0.02~0.05
power_6
0.16~0.30
power_3
0.05~0.09
power_7
0.30~1.00
power_4
0.09~0.13
power_all
0~1.00
Tab.2Names of frequency domain characteristic parameters of PPG signals and their meanings
Fig.5Standardized power spectrum based on 10 second PPG signal
Fig.6Impact of parameter values on feature selection results
序号
特征参数
序号
特征参数
1
FuzzyEn
10
$ {S_{OAO'}} $
2
RCMDE
11
power_all
3
$ {k_{O'A}} $
12
power_7
4
PermEn
13
$ {S_{OAM}} $
5
power_1
14
SampleEn
6
power_5
15
K值
7
power_4
16
$ {k_{OA}} $
8
power_6
17
t1
9
h
18
power_2
Tab.3Effective feature set constructed based on feature selection results
模型
A/%
SS/(obs·s?1)
Tinfer/s
S/kB
DT
40.0
~14000
5.16
573
KNN
96.0
~13000
0.63
517
Linear SVM
94.0
~15
1084.50
89547
QuadraticSVM
94.8
~12
1114.60
137104
Cubic SVM
94.8
~12
1120.50
131862
Gaussian SVM
97.6
~11
1101.50
159960
LDA
98.4
~12000
0.86
1048
Tab.4Different models of identity recognition
Fig.7Accuracy of KNN classification models under different neighbor numbers
核函数
k1
k2
A/%
SS(obs·s?1)
Linear
2.5
—
96.8
~14
Quadratic
2.0
—
95.6
~11
Cubic
2.0
—
95.2
~11
Gaussian
2.0
6.5
98.0
~11
Tab.5Optimization results of SVM model parameters
Fig.8Accuracy of SVM classification model before and after parameter optimization
Fig.9Classification accuracy of feature sets with different dimensions
文献
年份
M（PPG数据）
特征提取
分类器
A/%
注：1) FAR为假接受率；2) FRR为假拒绝率.
文献[30]
2003
17
时域
KNN
94.0
文献[9]
2003
17
时域
模糊逻辑
94.0
文献[5]
2014
40
时域
KNN
94.4
文献[6]
2015
10
时域
贝叶斯网络
97.5
文献[31]
2015
10（708组）
时域
FNN
FAR:4.21) FRR:3.72)
文献[7]
2016
15
时域
LDA
100.0
文献[10]
2016
23
高斯参数
LDA
95.7
文献[32]
2017
42
小波变换
KNN
99.8
文献[15]
2017
42
小波变换
SVM
100.0
文献[33]
2020
20
时域+频域
SVM
93.1
文献[34]
2021
100
—
CNN+LSTM
98.0
文献[35]
2021
35
频域
CNN
99.4
文献[12]
2022
100
时域+频域+小波变换
CNNLSTM
98.3
文献[11]
2023
50
—
MsNRPNet
92.0
本研究
25（125组）
时域+频域+非线性特征
LDA
100.0
—
50（250组）
99.9
200（1000组）
98.4
Tab.6Comparison of proposed PPG identity recognition with relevant research
[8]
SUN Bin, XUE Yunan, CHEN Xiaohui A method for PPG signal identification based on improved neighborhood rough set[J]. Foreign Electronic Measurement Technology, 2022, 41 (6): 8- 13
[9]
GU Y Y, ZHANG Y T. Photoplethysmographic authentication through fuzzy logic [C]// IEEE EMBS Asian-Pacific Conference on Biomedical Engineering . Kyoto: IEEE, 2003: 136–137.
[10]
SALANKE N S G R, MAHESWARI N, SAMRAJ A, et al. Enhancement in the design of biometric identification system based on photoplethysmography data [C]// International Conference on Green High Performance Computing . Nagercoil: IEEE, 2013: 1–6.
[11]
JEON Y J, KANG S J Multi-slice nested recurrence plot (msnrp): a robust approach for person identification using daily ECG or PPG signals[J]. Engineering Applications of Artificial Intelligence, 2023, 126: 106799
doi: 10.1016/j.engappai.2023.106799
[12]
LEE J A, KWAK K C Personal identification using an ensemble approach of 1D-LSTM and 2D-CNN with electrocardiogram signals[J]. Applied Sciences, 2022, 12 (5): 2692
doi: 10.3390/app12052692
[13]
ELGENDI M On the analysis of fingertip photoplethysmogram signals[J]. Current cardiology reviews, 2012, 8 (1): 14- 25
doi: 10.2174/157340312801215782
[14]
HAYES M J, SMITH P R A new method for pulse oximetry possessing inherent insensitivity toartifact[J]. IEEE Transactions on Biomedical Engineering, 2001, 48 (4): 452- 461
doi: 10.1109/10.915711
[15]
KARIMIAN N, GUO Z, TEHRANIPOOR M, et al. Human recognition from photoplethysmography (PPG) based on non-fiducial features [C]// IEEE International Conference on Acoustics, Speech and Signal Processing . New Orleans: IEEE, 2017: 4636–4640.
[16]
ALAM R-U, ZHAO H, GOODWIN A, et al Differences in power spectral densities and phase quantities due to processing of EEG signals[J]. Sensors, 2020, 20: 6285
doi: 10.3390/s20216285
[17]
HU J F. Biometric system based on EEG signals: a nonlinear model approach [C]// International Conference on Machine Vision and Human-machine Interface . Kaifeng: IEEE, 2010: 48–51.
[18]
HEJAZI M, AL-HADDAD S A R, HASHIM S J, et al. Non-fiducial based ECG biometric authentication using one-class Support Vector Machine [C]// Signal Processing: Algorithms, Architectures, Arrangements, and Applications . Poznan: IEEE, 2017: 190–194.
[19]
YE Z, TURNER R. Intelligent linear and nonlinear analysis for biometric fingerprint recognition [C]// 39th Southeastern Symposium on System Theory . Macon: IEEE, 2007: 315–319.
[1]
RAJARAM S, VOLLALA S, RAMASUBRAMANIAN N, et al Enhanced and secured random number generation for eUASBP[J]. International Journal of System Assurance Engineering and Management, 2022, 13 (3): 1135- 1150
[2]
SPOOREN J, PREUVENEERS D, JOOSEN W. PPG2Live: using dual PPG for active authentication and liveness detection [C]// International Conference on Biometrics . Crete: IEEE, 2019: 1–6.
[20]
LEE T K M, SANEI S, CLARKE B. Fusion of nonlinear measures in fronto-normal gait recognition [C]// 5th International Multi-conference on Computing in the Global Information Technology . Valencia: IEEE, 2010: 104–109.
[21]
PINCUS S M Approximate entropy as a measure of system complexity[J]. Proceedings of the National Academy of Sciences, 1991, 88 (6): 2297- 2301
doi: 10.1073/pnas.88.6.2297
[3]
NOWARA E M, SABHARWAL A, VEERARAGHAVAN A, et al. PPG secure: biometric presentation attack detection using photopletysmograms [C]// IEEE International Conference on Automatic Face and Gesture Recognition . Washington, DC: IEEE, 2017: 56–62.
[4]
GU Y Y, ZHANG Y, ZHANG Y T. A novel biometric approach in human verification by photoplethysmographic signals [C]// 4th International IEEE EMBS Special Topic Conference on Information Technology Applications in Biomedicine . Birmingham: IEEE, 2003: 13–14.
[22]
章思佳, 林栋, 齐诗仪, 等 基于眼电信号近似熵探讨针灸治神的临床应用[J]. 中国针灸, 2023, 43 (1): 79- 82 ZHANG Sijia, LIN Dong, QI Shiyi, et al Clinical application of acupuncture-moxibustion for the treatment of spirit based on approximate entropy of electrooculogram signal[J]. Chinese Acupuncture and Moxibustion, 2023, 43 (1): 79- 82
[23]
樊凤杰, 白洋, 纪会芳 基于脑电非线性动力学特征探究TEAS内关穴对焦虑的影响[J]. 计量学报, 2022, 43 (1): 133- 139 FAN Fengjie, BAI Yang, JI Huifang Explore the lnfluence of transcutaneous electrical acupoint stimulation neiguan point on anxiety based on electroencephalogram nonlinear dynamic analysis method[J]. Acta Metrologica Sinica, 2022, 43 (1): 133- 139
doi: 10.3969/j.issn.1000-1158.2022.01.21
[5]
KAVSAOGLU A R, POLAT K, BOZKURT M R A novel feature ranking algorithm for biometric recognition with PPG signals[J]. Computers in Biology and Medicine, 2014, 49: 1- 14
doi: 10.1016/j.compbiomed.2014.03.005
[6]
JAAFAR N A L, SIDEK K A, AZAM S N A M. Acceleration plethysmogram based biometric identification [C]// International Conference on BioSignal Analysis, Processing and Systems . Kuala Lumpur: IEEE, 2015: 16–21.
[7]
CHAKRABORTY S, PAL S. Photoplethysmogram signal based biometric recognition using linear discriminant classifier [C]// International Conference on Control, Instrumentation, Energy and Communication . Kolkata: IEEE, 2016: 183–187.
RICHMAN J S, MOORMAN J R Physiological time-series analysis using approximate entropy and sample entropy[J]. American Journal of Physiology-Heart and Circulatory Physiology, 2000, 278 (6): H2039- H2049
doi: 10.1152/ajpheart.2000.278.6.H2039
[25]
CHEN W, WANG Z, XIE H, et al Characterization of surface EMG signal based on fuzzy entropy[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2007, 15 (2): 266- 272
doi: 10.1109/TNSRE.2007.897025
[26]
AZAMI H, ROSTAGHI M, ABÁSOLO D, et al Refined composite multiscale dispersion entropy and its application to biomedical signals[J]. IEEE Transactions on Biomedical Engineering, 2017, 64 (12): 2872- 2879
doi: 10.1109/TBME.2017.2679136
徐慧, 付迎春, 付朝川, 等 改进WOA算法优化SVM的网络入侵检测[J]. 实验室研究与探索, 2019, 38 (8): 128- 133 XU Hui, FU Yingchun, FU Chaochuan, et al Improved whale optimization algorithm to optimize support vector machine for network intrusion detection[J]. Research and Exploration in Laboratory, 2019, 38 (8): 128- 133
doi: 10.3969/j.issn.1006-7167.2019.08.031
[29]
MOODY G B, MARK R G. A database to support development and evaluation of intelligent intensive care monitoring [C]// Computers in Cardiology . Indianapolis: IEEE, 1996: 657–660.
[30]
GU Y Y, ZHANG Y, ZHANG Y T. A novel biometric approach in human verification by photoplethysmo-signals [C]// 4th International IEEE EMBS Special Topic Conference on Information Technology Applications in Biomedicine . Birmingham: IEEE, 2003: 13–14.
[31]
LEE A, KIM Y. Photoplethysmography as a form of biometric authentication [C]// 2015 IEEE Sensors . Busan: IEEE, 2015: 1–2.
[32]
KARIMIAN N, TEHRANIPOOR M, FORTE D. Non-fiducial PPG-based authentication for healthcare application [C]// IEEE EMBS International Conference on Biomedical and Health Informatics . Orlando: IEEE, 2017: 429–432.
[33]
KHAN M U, AZIZ S, NAQVI S Z H, et al. Pattern analysis towards human verification using photoplethysmograph signals [C]// International Conference on Emerging Trends in Smart Technologies . Karachi: IEEE, 2020: 1–6.
[34]
HWANG D Y, TAHA B, LEE D S, et al Evaluation of the time stability and uniqueness in PPG-based biometric system[J]. IEEE Transactions on Information Forensics and Security, 2021, 16: 116- 130
doi: 10.1109/TIFS.2020.3006313
